Krashen through the eyes of Cognitive Load Theory

The problem with Krashen - What about i = 0

When Krashen focuses on comprehensible input, it seems that he implicitly focuses on the advanced user. He focuses on the progression from i to i + 1, but this becomes difficult when i = 0.

However,for a total novice, this might be too complex. The user would be confronted with many elements at once because only together they convey the meaning that Krashen emphasizes so heavily. Krashen's whole point of meaning is only possible, because words have such a high level interactivity. He emphasizes, that this meaning cannot be learned through (learned) theory, but only from comprehensible input.

For a novice, however, CLT predicts, that this will inevitably lead to a cognitive overload for the user. It is thus unlikely for input to be comprehensive for him, because the interactivity of words is far too high. For input to be comprehensible to a novice, it would need to be in the form of single words.

CLT's solution - simplify the language schema

CLT's solution to this, as stated above in the CLT section 5, is to remove the interactivity of elements. This will lead to the acquisition of an incorrect or at least incomplete schema.

Applied to language, words can be viewed to have individual meanings but also combined meanings because a sequence of words means that their individual meanings are combined into the meaning of the sequence. This sequence meaning can also differ, depending on the order, a slight variation within the words, and the context of the sequence.

This means to initially ignore the interactivity of words and learn them in a highly simplified form. This is where, in addition to the language schema, a second schema - the word schema - comes into play: The language schema consists of many elements, which are words. In Krashen's terms this can be expressed as the input being, more or less, the language schema, and the number of elements (words) it consists of which are Krashen's i. Those words have many relationships and interactions with each other. To improve the language schema, more and more elements have to be added. A schema can, however, also consist of sub-schemas, which means that an element of a schema can also be a schema. In this case the sub-schema is, in contrast to the language schema, a word schema. Because we need to remove the interactions and complexity, the word schema would start as an extremely simple schema consistent of only 2 elements: the word in the base language, the L1 language, and that translation in the target language, the L2 language. For the word 'cat', the schema would consist of the elements 'cat' and 'gato' (when Spanish is the L2).

This extremely simple schema is far from the requirements of language acquisition, which seeks to understand the nuances of communicative language. It is, however, in accordance with the need for simplified schemas. CLT now proceeds to automate and add to these word schemas. This means adding more elements, such as the determiner, the gender or relations the word schema could have with other word schemas, for example the verb 'running' could be used in connection to the word schema 'cat'. This is in line with CLT as not only the elements and schemas of CLT are unique to every user, that there is also no clear boundary for how simple or complex a schema is and where an element ends and where a schema starts.

While at the beginning, as CLT predicts, the user has to use a lot of mental energy, over time he will automate the word schema, resulting more and more in smooth and flawless schema usage in form of recalling word translations, as in that 'gato' means 'cat'. It then becomes possible for the user to acquire more complex schemas.

The role of the language and word schema in Krashen's theory

This focus on practicing word schemas in order to automate them seemingly contradicts Krashen, who states, that the conscious nature of learning is not transferable to the unconscious process of acquisition. However, as follows, it can be integrated to supplement his theory.

The language schema is not concerned with word translations or rules, but rather with word meanings and relationships, as well as using the elements (words) in their combination, to extract meaning. As individual letters form a meaningful word, individual words form a meaningful expression.

If we separate the language form the word schema, it becomes clear, that they follow different goals. The word schema provides a mental dictionary (or word bank) to the user and the language schema provides the user with the ability to actually use this word bank for communication. Thus, for the language schema, Krashen holds true. This is not the case for the word schema. However, Krashen doesn't really refer to it in the first place, instead he refers to and focuses on the language schema. This independence of the two schemas, however, is, according to CLT, not appropriate. Instead, so CLT, the simplified word schema serves as a scaffold to the complex language schema.

The strong point of Krashen's 'acquisition' comes after the user has become more and more proficient in his initial bank of word schemas. The intrinsic cognitive load of the language schema will be progressively reduced, because the individual meanings of the words are known, and the user can focus on the combined meanings. In the words of CLT, the user has acquired the oversimplified schemas and can now better learn the full real schema with all its complexity and interdependence.

The transition from word to language, from learning to acquisition

It follows, that the building and automation of word schemas becomes less important now, instead a shift of focus to the language schema becomes optimal. In this sense, the more word schemas, the easier it becomes to develop the language schema. The better the language schema, the easier it likely becomes to add new word schemas as elements into the language schema with input. This circular practice deepens and expands the language schema and leads to its automation. This now is not only in line with Krashen, but also supports his learning acquisition-distinction as follows: Should the user only learn and automate the word schema, which are an artificial simplification of the language schema, he will only learn this 'untrue' simplified schema.

As a consequence, if the user has learned and automated a high amount of word schemas, eg. 5000, but has not advanced to the language schema, he has a high degree of competency in using his mental dictionary, but he is unable to apply them for communication. For example, being an expert in the alphabet, knowing every letter and its pronunciation, and maybe even its most likely neighboring letter, but having no schemas of words. What good is perfect knowledge of the alphabet, if no words are known? After having learned the alphabet, now that the user has learned many words, it also becomes easy to learn new letters. If, for some reason the user has to learn new letters, such as German umlaute or French accents, as might be the case in other languages than English, learning these likely becomes very easy, and the same energy required to initially learn the alphabet is not needed anymore. The same applies to L2 acquisition. When the focus has shifted to automating the language schema, the user is somewhat proficient in comprehending (reading and understanding) input, and a new word can be picked up much more easily.

In addition, old words with their respective schemas, are further reviewed, modified if needed (with a new context of a word) and automatized. Also, elements (words) of the language schema are less viewed as word schemas, than as elements, which follow Krashen's learning-acquisition dichotomy. For example, when reading a word, no conscious attention to the letters is given, instead they are read and combined into words subconsciously. In this sense, when reading the word 'gato', this is not seen as a word schema, that now needs to be translated into cat, but instead, 'gato' simply means 'cat', as in 'gato' is just another label for the animal cat. At this schema development stage, comprehensible input has become the best way of automating the language schema.

Conclusion on comparing Krashen with CLT

It is, however, unclear if a simplified word schema and a real word schema, that is an element of the language schema, can be the same thing. The question is: If a simple word schema is sufficiently extended and automatized, will it become a word schema, that is part of the language schema, or in other words: Can a simple word schema be transferred into a real word schema? This would mean that learning a simplified word schema would directly improve the language schema. Or can a simple word schema not be transferred, but instead it only helps in building real world schemas for the language schema? In that case, learning a simplified word schema would not benefit the language schema, but instead reduce the extraneous load of building the language schema by input, because the user can look up the word in his mind.

An argument against the direct transfer would be, if a user always has to 'translate' (in his mind) his learned simplified word schemas, but in contrast he 'just knows' real word schemas. If this were the case, it would be in line with Krashen's learning, where a user can only use learned material with monitor use. An argument for the direct transfer would be if the user, with enough automation of the simple word schema, is able to 'just know' (acquire) elements.

While there is no clear answer and this is just speculation, maybe some elements of a simple word schema can be transferred to a certain degree, while other elements can be less so. Maybe this would be with variation to ease, speed and other factors. This could mean elements such as translations of words might be transferable, while elements such as certain grammatical rules might be much less so and would have to be acquired through input (via the language schema) as indicated by Krashen (Krashen, 1982, p. 79ff, p. 105ff).

Ultimately though, this remains unanswered, but it seems plausible, that simple word schemas can at least aid with real world schemas. It would explain, why Krashen suggests learned grammar cannot be acquired but critics of Krashen claim, that deliberately learned words can be acquired.

While CLT confirms Krashen's theory, it also indicates the potential of learning, practice, and drill, which appear to be the optimal way of word schema automation. It remains unanswered, where exactly the transfer happens from word schema automation to language schema automation.

It also seems in regard to motivation, that a fast transfer or start of focus of cognitive load to language schema automation is optimal, because the word schema are just a help, a means to an end, and thus not the reason for learning a language in the first place. As a consequence, it seems a plausible goal to start with language schema acquisition as soon as possible. Therefore, we next see if 'learning' is as efficient as our insights, from Cognitive Load Theory, would suggest.